feat(train-scheduling): implement day-level booking pool

- Added `unplaced` method in `BookingNotifierService` to log warnings for bookings that cannot be placed on any train.
- Introduced `getAvailableDays` method in `TrainSchedulingService` to retrieve distinct days with open departures for a given route.
- Created `AvailableDaysQueryDto` for querying available days based on origin and destination yards.
- Updated `TrainSchedulingController` to expose an endpoint for available days.
- Modified frontend components to support day-level booking, allowing customers to select only a day without pinning to a specific train.
- Removed references to train schedules in booking forms and review steps, emphasizing day selection.
- Added a database migration to create an index for efficient querying of bookings by route and day.

apps/edr-freight-api/src/modules/bookings/entities/booking.entity.ts

@@ -277,7 +277,15 @@ export class Booking extends BaseEntity {
   @Column({ name: 'scheduled_at', type: 'timestamptz', nullable: true })
   scheduledAt?: Date | null;
 
-  /** The schedule this booking targets (pool membership), set at creation. FK to train_schedules. */
+  /**
+   * The train this booking is assigned to. FK to train_schedules.
+   *
+   * Day-level pooling: customers no longer pick a train — they pick a DAY, and
+   * this stays null at creation. The batch engine sets it when it assigns the
+   * booking to a specific train within its (route, day) pool; staff may also
+   * pin it manually. The day-level pool is keyed on
+   * (origin_yard_id, destination_yard_id, day of scheduled_date), not this column.
+   */
   @Column({ name: 'train_schedule_id', type: 'uuid', nullable: true })
   trainScheduleId?: string | null;